Agent Onboarding

Everything an AI agent needs to start using GTM Tools. There are three ways in: MCP, REST, and CLI.

Integration paths

1. MCP server (recommended)

Each server exposes its tools over MCP Streamable HTTP. Point your MCP client at the relevant subdomain with your bearer token.

Add only the servers your agent needs — there’s no penalty for omitting one.

Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

1
{
2
  "mcpServers": {
3
    "gtm-tools": {
4
      "url": "https://api.gtm-tools.sh/mcp",
5
      "headers": { "Authorization": "Bearer YOUR_KEY" }
6
    },
7
    "gtm-tools": {
8
      "url": "https://api.gtm-tools.sh/mcp",
9
      "headers": { "Authorization": "Bearer YOUR_KEY" }
10
    },
11
    "gtm-tools": {
12
      "url": "https://api.gtm-tools.sh/mcp",
13
      "headers": { "Authorization": "Bearer YOUR_KEY" }
14
    },
15
    "gtm-tools": {
16
      "url": "https://api.gtm-tools.sh/mcp",
17
      "headers": { "Authorization": "Bearer YOUR_KEY" }
18
    }
19
  }
20
}

Restart Claude Desktop. Tools should appear in the connector panel.

Cursor

Add to .cursor/mcp.json in your project root:

1
{
2
  "mcpServers": {
3
    "gtm-tools": {
4
      "url": "https://api.gtm-tools.sh/mcp",
5
      "headers": { "Authorization": "Bearer YOUR_KEY" }
6
    },
7
    "gtm-tools": {
8
      "url": "https://api.gtm-tools.sh/mcp",
9
      "headers": { "Authorization": "Bearer YOUR_KEY" }
10
    }
11
  }
12
}

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

1
{
2
  "mcpServers": {
3
    "gtm-tools": {
4
      "url": "https://api.gtm-tools.sh/mcp",
5
      "headers": { "Authorization": "Bearer YOUR_KEY" }
6
    }
7
  }
8
}

2. REST API

Every tool is a POST /api/v0/{tool_name} on its server’s subdomain. Same JSON body, same auth header. See the REST API page for the full pattern.

$
curl -X POST https://api.gtm-tools.sh/api/v0/get_linkedin_company_url \
>
  -H "Authorization: Bearer $GTM_TOOLS_API_KEY" \
>
  -H "Content-Type: application/json" \
>
  -d '{"domain": "siena.cx"}'

Tool discovery: GET /api/v0 on any server returns its tool index.

$
curl https://api.gtm-tools.sh/api/v0 \
>
  -H "Authorization: Bearer $GTM_TOOLS_API_KEY"

3. CLI

Install once, point at all four servers.

$
npm install -g gtm-tools
$
gtm-tools admin balance
$
gtm-tools socials company-url siena.cx
$
gtm-tools signals detect gymshark.com

See the CLI page for the full command reference.

Available MCP tools by server

Admin Tools

Socials Tools

Email Tools

Signals Tools

Quick start code

End-to-end prospect research

1
const KEY = process.env.GTM_TOOLS_API_KEY!;
2
const HEADERS = {
3
  "Authorization": `Bearer ${KEY}`,
4
  "Content-Type": "application/json",
5
};
6
7
async function call<T>(tool: string, body: unknown): Promise<T> {
8
  const res = await fetch(`https://api.gtm-tools.sh/api/v0/${tool}`, {
9
    method: "POST",
10
    headers: HEADERS,
11
    body: JSON.stringify(body),
12
  });
13
  if (!res.ok) throw new Error(`${tool}: ${res.status} ${await res.text()}`);
14
  return res.json();
15
}
16
17
// 1. Detect signals to qualify the account (free dispatch + 5/detector)
18
const signals = await call<any>("detect_signal", {
19
  domain: "gymshark.com",
20
});
21
22
// 2. Find decision-makers (30 tokens)
23
const employees = await call<any>("list_linkedin_company_employees", {
24
  domain: "gymshark.com",
25
  title_filters: "(VP OR Director OR Head) AND (CX OR Support) NOT intern",
26
  limit: 10,
27
});
28
29
// 3. Verify professional emails (5 tokens each)
30
for (const e of employees.results.slice(0, 3)) {
31
  const { email, is_catch_all } = await call<any>("get_email", {
32
    name: e.name,
33
    domain: "gymshark.com",
34
  });
35
  console.log(e.name, "→", email, is_catch_all ? "(catch-all)" : "");
36
}

Documentation for agents

Add to your agent’s system prompt:

For GTM Tools usage, refer to: https://gtm-tools.sh/llms.txt

Next steps